Acute-on-chronic liver failure (ACLF) is a severe complication of cirrhosis characterized by multiorgan failure and high short-term mortality. Current diagnostic criteria, such as the EASL-CLIF organ failure score, have limitations in capturing the full spectrum of organ dysfunction and stratifying risk. In a new international study published in the Journal of Hepatology, researchers developed and validated the A-TANGO organ failure (OF) score to address these gaps and improve prognostic accuracy across diverse populations.
The A-TANGO OF score redefines organ failure thresholds and adds a new ACLF grade 4 to better capture high-risk patients.
Compared to the CLIF-C OF score, A-TANGO increased ACLF diagnosis from 24% to 36% in the derivation cohort while preserving predictive accuracy for 28-day and 90-day mortality.
The score demonstrated an integrated discrimination improvement (IDI) of 16.2% and a net reclassification improvement (NRI) of 16.1% over the CLIF-C OF score.
Findings were externally validated in large cohorts from India and China, supporting global applicability.
Study Snapshot
Design: Retrospective analysis of prospective observational cohorts
Derivation cohort: 3,896 patients from European and Latin American studies (CANONIC, PREDICT, ACLARA)
Validation cohorts: 2,055 patients from India (Ambi-spective study) and 2,568 patients from China (CATCH-LIFE)
Study period: 2011–2023, follow-up completed 2023
Primary outcome: 28-day and 90-day mortality
Key comparison: A-TANGO OF score vs. CLIF-C OF score
Why This Study Matters
ACLF carries a high risk of death, and accurate risk stratification is essential for clinical decision-making and trial design. The EASL-CLIF criteria are widely used but have been criticized for not capturing all organ dysfunctions and for having wide mortality variation within the highest grade. The A-TANGO OF score was designed to refine organ failure definitions and add a grade 4 ACLF to better identify patients at extreme risk. This could lead to more precise prognostication and more efficient clinical trials.
How the Study Was Conducted
The investigators used data from three large EF-CLIF consortium studies (CANONIC, PREDICT, ACLARA) in Europe and Latin America to derive the A-TANGO OF score. They defined organ failure for each of six organ systems using three subscores, with subscore 3 corresponding to a ≥15% 28-day mortality risk. The score then generates an overall ACLF grade (1–4). The new score was externally validated in an Indian ambi-spective cohort and a Chinese prospective cohort (CATCH-LIFE). Prognostic performance was compared with the CLIF-C OF score and the MELD-Na score using C-index, integrated discrimination improvement (IDI), net reclassification improvement (NRI), Brier score, and decision curve analysis. Additionally, two simpler prognostic models (A-TANGO ACLF-WBC and A-TANGO ACLF-CRP) were derived from the score.
What the Researchers Found
In the derivation cohort, the A-TANGO OF score identified 1,305 patients (36%) with ACLF, compared to 863 patients (24%) using the CLIF-C OF score (p <0.001). Patients diagnosed with ACLF by A-TANGO had lower mortality than those diagnosed by CLIF-C (28-day: 30% vs. 36%, p=0.0014; 90-day: 40% vs. 48%, p=0.0007), reflecting the inclusion of less severely ill patients at lower mortality risk. Of the 430 patients newly identified by A-TANGO (mostly grade 1), 28-day mortality was 15.6%. The increase in ACLF diagnosis was driven mainly by more frequent identification of coagulation, brain, respiratory, and kidney failure.
Fig. 1 Cumulative incidence of death according to ACLF grade based on the A-TANGO score or the CLIF-C OF score.
The A-TANGO OF score showed strong correlation with observed mortality (R²=0.94 for 28-day, 0.97 for 90-day) and a per-unit hazard ratio of 1.48 (28-day) and 1.37 (90-day). The C-index for 28-day mortality was 0.79, similar to 0.78 for CLIF-C OF and MELD-Na. While the area under the receiver operating characteristic curve did not improve, the IDI of 0.162 (16.2%) and NRI of 16.1% indicated meaningful improvement in reclassification and discrimination across the risk spectrum. Decision curve analysis showed similar clinical utility for both scores. The A-TANGO ACLF-WBC and ACLF-CRP models also showed strong prognostic performance. All findings were confirmed in the Indian and Chinese validation cohorts.
What the Findings May Mean
The A-TANGO OF score offers a more sensitive tool for diagnosing ACLF without sacrificing predictive accuracy. By capturing more patients with organ failure, especially at lower grades, the score may identify individuals who could benefit from early intervention. The addition of grade 4 addresses a gap in the current criteria for the highest-risk patients. The improved risk classification (NRI 16%) and discrimination (IDI 16%) suggest that A-TANGO provides a more clinically meaningful stratification. However, the authors caution that these improvements do not necessarily translate to better clinical outcomes unless coupled with effective treatments.
Strengths and Limitations
Strengths include the large multinational derivation and validation cohorts (over 8,500 patients), prospective data collection, and comprehensive statistical evaluation including net reclassification and decision curve analysis. The study also provides simpler bedside models (ACLF-WBC, ACLF-CRP) that may be easier to implement.
Limitations include the retrospective analysis of prospective data, missing data, differences in cohort recruitment periods (2011–2023), and changes in clinical practice over time. The use of historical classification criteria for some cohorts may affect consistency. The authors also note that the increase in ACLF diagnosis may include patients with less severe disease, which could affect trial endpoint interpretations. External validation in regions with different etiologies and healthcare settings is still needed.
Implications for Practice and Research
The A-TANGO OF score provides a reproducible and comprehensive framework for diagnosing ACLF and assessing treatment response. Its ability to identify more patients with organ failure may reduce sample size requirements in clinical trials by enriching the study population with events. The score can also serve as a dynamic endpoint, such as ACLF resolution, for evaluating emerging therapies. In clinical practice, the score and its simpler variants can help guide treatment decisions and prognosis discussions with patients and families.
References
Engelmann C, Verma N, Qi T, Kerbert AJC, et al. Development and validation of the A-TANGO organ failure score for acute-on-chronic liver failure in global cohorts. J Hepatol. 2026 Jul;85(1):91-105. doi: 10.1016/j.jhep.2026.02.017. PMID: 41730386.
